Jason Langley1, Daniel E Huddleston2, Michael Merritt1, Xiangchuan Chen1, Rebecca McMurray2, Michael Silver2, Stewart A Factor2, and Xiaoping Hu1
1Department of Biomedical Engineering, Emory University & Georgia Tech, Atlanta, GA, United States, 2Department of Neurology, Emory University, Atlanta, GA, United States
Synopsis
Inconclusive results from prior diffusion tensor
imaging-based studies can be attributed to variability in location of regions
of interest used to define the substantia nigra and its subcomponents. We apply recent findings from neuromelanin sensitive MRI to
standardize regions of interest for the substantia nigra. Differences in fractional anisotropy and mean diffusivity were found in the neuromelanin sensitive substantia nigra but not in the substantia nigra defined in the b0 image.Introduction
Multiple
studies have investigated Parkinson’s disease (PD) related changes in fractional
anisotropy (FA) and mean diffusivity (MD) in the substantia nigra (SN). While some
of these studies revealed lower FA in the SN for PD groups1-4, others did
not find differences in FA between PD and controls in the SN regions of
interest (ROIs)5-7. The discrepancy can be attributed to variability in the
anatomic location of ROIs used to define the SN across studies. In those
studies, ROIs manually drawn on T2 weighted (T2w) spin echo images could be
highly subjective and variable. This variability has been a confounding factor
in identifying potential PD biomarkers from DTI metrics (illustrated in Figure 1)7.
In
addition, recent work has shown that the SN seen in T2w and neuromelanin
sensitive MRI (NM-MRI) contrasts are spatially incongruent8 and this incongruence
could account for the wide range of results in previous studies (Figure 2)1-7. In the present work, we remove the ROI
variability in previous studies by creating standardized SN pars compacta (SNpc)
ROIs, defined in NM-MRI and SWI images from healthy controls and denoted NM-MRI SNpc
ROI and T2w SN ROI, respectively. These ROIs are then used to investigate alterations in diffusion measures,
namely MD and FA, due to PD.
Methods
A cohort
of 37 subjects, who provided written, IRB-approved, informed consent, were
studied. Demographic data for the cohort is summarized in Table 1.
All data
were acquired on two 3 T MRI scanners (TRIO, Siemens Medical Solutions,
Malvern, PA) at Emory University using a 12 channel receive only coil. Images
from an MP-RAGE sequence (echo time (TE)/repetition time (TR)/inversion time=3.02/2600/800 ms, flip angle=8°, voxel size=1.0×1.0×1.0 mm3) were used for registration
from subject space to common space. Diffusion MRI data were collected with a
single-shot spin-echo, EPI sequence. Diffusion-weighting gradients were applied
in 64 directions with a b value of 1000 s/mm2; TE/TR=97/3292 ms,
FOV=212×212 mm2, matrix size of 106×106, voxel size=2×2×2 mm3,
64 slices with no gap, covering the entire brain. Two sets of
diffusion-weighted images, with phase-encoding directions of opposite polarity,
were acquired to correct for susceptibility distortion9. For each
diffusion-weighted acquisition, six images without diffusion weighting (b0
images) were also acquired. The processing pipeline for DTI and common space
registration are shown in Figure 3. Standard space
NM-MRI SNpc and T2w ROIs from literature8 were used as ROIs for the NM-MRI SNpc ROI and T2 SN ROI, respectively.
Results
Mean FA values for the NM-MRI SNpc ROI were lower
in the PD group than the control group (p=0.001).
In contrast, no statistically significant difference in mean FA was seen in the
T2w SN ROI (p=0.36). In addition, the
average mean diffusivity (MD) offered no statistically significant difference
between the two groups in the SN for the T2w SN ROI (p=0.43), while a statistically significant difference was seen in
mean MD between PD and control groups for the NM-MRI SNpc ROI(p=0.01). These results are summarized in
Figure 4.
The side more affected by PD in each subject by
determined by comparing the lateralized sums of UPDRS-III subscores. In the SN
volume contralateral to the more affected side, the PD group showed lower FA
than the control group (PD: 0.36±0.03; CO: 0.41±0.03; p<10-4). In addition, a lateralized SN degeneration in
the NM-MRI SNpc ROI consistent with the disease laterality was seen.
Specifically, greater disease effects were seen in the FA in the rostral
portion of the SN on the side contralateral to the more affected than the other
side (p=0.03).
Discussion
The inconsistency in iron deposition as well as the location and volume
of SN ROI, as shown in the b0 map, could explain the discrepancy between
studies showing significant changes in the SN from PD [1-4] and those that
found no difference in the SN [5-7]. All DTI-based studies [1-7] used a similar
definition for SN ROIs [4]. In controls, we found little overlap in the
hypointense region in the b0 map and the NM-MRI SNpc ROI, consistent with a
recent report that found the overlap between the NM-MRI SNpc ROI and the T2w
hypointense SN ROI to be ~10% in normal subjects [8]. PD patients with less
iron deposition and less overlap between NM-MRI SNpc and T2w hypointense SN ROIs
could appear to have a smaller disease effect on DTI measures as the ROIs would
primarily be placed in the T2w SN ROI, which was found to exhibit no
disease-related changes in DTI.
Acknowledgements
This work was partially supported by the Michael J. Fox Foundation (MJF 10854) and NINDS Parkinson's Disease Biomarkers Program U18
Award (U18 NS082143).References
1. Chan, LL, et al. J Neurol Neurosurg Psychiatry 2007; 78(12):1383-6.
2. Du G, et al. Mov Disord 2011; 26(9): 1627-32.
3. Peran P,
et al. Brain 2010; 133(11): 3423-33
4. Vaillancourt
DE, et al. Neurology 2009;
72(16): 1378-84.
5. Aquino D, et al. Neurol Sci 2014; 35(5): 753-8.
6. Menke RA, et al. NeuroImage 2009; 47(2): 435-41.
7. Schwarz ST, et al. Neuroimage Clin 2013; 3: 481-8.
8. Langley J, et al. NeuroImage 2015; 112(1): 7-13
9. Andersson
JL, et al.
NeuroImage 2003; 20(2): 870-88